recipe-plan
Pass
Audited by Gen Agent Trust Hub on Nov 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implements a legitimate workflow for project orchestration. It coordinates between specialized agents to transform design requirements into actionable plans. All actions are confined to standard documentation paths, and no suspicious command execution or data exfiltration patterns were identified.- [INDIRECT_PROMPT_INJECTION]: The skill's architecture includes an ingestion surface for external project files, which creates a potential vector for indirect instructions.
- Ingestion points: File content from
docs/design/and PRD documents (SKILL.md, Steps 1-4). - Boundary markers: Instructions specify mechanical extraction into fields but do not define specific escaping or protective delimiters for the file content.
- Capability inventory: The workflow involves file-writing capabilities through sub-agents like
work-planner. - Sanitization: No explicit content validation or sanitization is performed on the ingested design docs.
- Contextual note: The inclusion of a mandatory user approval step in Step 5 serves as a primary control to mitigate the impact of potentially malicious content embedded in design documents.
Audit Metadata